My July 13th lecture seems to have been challenging for people. This was, I think, the most conceptually difficult material we will deal with in this class, but I still fear my lecture may have been a bit too improvised and not adequate to bring complex concepts within reach of undergraduates. This blog post is an attempt to make up for that. I will use the charts I presented in class and make another attempt to explain.
First, let us make a careful consideration of the standard supply and demand curve:
In this version of supply and demand-- probably the most familiar chart in economics-- demand is downward-sloping and supply is upward-sloping. This model is especially well-suited to agricultural markets, where the marginal utility of any particular agricultural good falls because people get satiated with it, and where the marginal cost of producing an extra unit rises because land is scarce and farmers have to cultivate less fertile land.
But in most markets an objection can be made to the shape of the supply curve on the basis of the so-called "Law of Science." Put simply, it is an assumption of the natural sciences that "if you do the same thing twice, the same thing should happen." Applying this to, say, the widget industry, if you can produce one widget for $1, you should be able at least to repeat the process and produce another widget for $1, or 1 million widgets for $1 million. And you should be willing to sell them for $1, or, if not, it's not clear at any rate why you should demand a higher markup for the millionth widget than for the first. The "Law of Science" suggests, then, that supply curves should not be upward-sloping, unless some factor of production is fundamentally scarce. Of course, if your production process requires some special input-- if you are producing airships, for example, and you require helium-- and there is a global scarcity of helium (or if marginal helium sources are increasingly difficult to tap), then you can only supply more airships for a higher price. But it seems likely that most industries do not comprise such a large share of the demand for their basic inputs that they would drive up the price. In the short run, supply curves may still slope upwards, since it takes time to mobilize labor and, especially, to invest in capital. But long-run supply curves should be flat, as in Figure 2:
It should be borne in mind, when considering Figure 2, that the demand curve represents industry demand. It does not represent the demand curve faced by a single firm. A single firm which faced a demand curve like that in Figure 2 might be able to increasing revenues, and/or profits, by raising the price above P*. Indeed, if P*=MC=AC (if P* equals marginal cost equals average costs), which is the usual assumption underlying a supply curve such as that drawn in Figure 2, the firm will certainly not wish to charge P*. Instead, individual firms are assumed to be "price takers," charging the price that equilibrates supply and demand. But how can it be rational for firms to charge P*? Because there are many firms, and any firm that charges more than P* will see all its customers flee to its competitors. In other words, while industry demand is downward-sloping, each firm faces a flat demand curve, as shown in Figure 3:
It is instructive to consider the profit maximization problem of the individual firm. Under the price-taking assumption, the firm can sell nothing for a price greater than the market price, but at the market price it can sell whatever it offers for sale. The firm's cost structure of production is best represented by a marginal cost (MC) curve. Wherever MC is below D, the marginal cost of producing another unit is less than what that unit will sell for (which in this case is the same as the marginal revenue). In this case, the MC curve is U-shaped: when production volumes are low, MC is falling, but later it starts rising. The point where MC crosses D is the optimal quantity (Q*) for the firm to produce, where the last unit produced sells for only just enough to cover its production cost. (Note also that for the zero-profit condition to hold in the market as a whole (which it must if entry is not to occur), the area above MC and below D to the right of Q* must be equal to the area below MC and above D to the right of Q*.)
Unfortunately for the economic theorist, the existence of an optimal quantity for the firm to produce depends on the MC curve being U-shaped. It is not only logically possible but quite realistic that an MC curve might, instead, be falling forever, as in Figure 4:
Figure 4, please note, is an impossibility, an absurdity, but the reasons why are subtle and important. Let's stop and thinking what Figure 4 is saying. First, it is saying that a firm faces a flat demand curve: no matter how much or little it produces, it can sell that amount for the market price, but not a penny more. Second, it is saying that the more the firm produces, the lower are its marginal costs: that is, there are economies of scale in production. But how much will the firm choose to produce? Infinity! Every unit the firm produces has a lower cost than the last, but gets the same price in the market. More production means more profits, ad infinitum.
Clearly there is something wrong here, but what? It is certainly possible for MC to be falling, that is, for economies of scale to exist, over an arbitrarily large range of production scales. But if MC is falling over the whole relevant range, there cannot be perfect competition. Firms will become large enough that firm demand curves start to look like industry demand curves, sloping downwards. Indeed, an industry with permanently falling marginal cost is the theoretical paradigm case of natural monopoly.
Is monopoly a common phenomenon? This is a difficult question, not only to answer empirically, but even to define adequately. After all, outside agricultural markets, almost everything sold today is a bit different from even its closest substitutes (e.g., Duracell vs. Energizer batteries, Coke vs. Pepsi, Colgate toothpaste vs. CVS's store brand). The gas station at this corner of the street is not a perfect substitute for the gas station at that corner of the street, given traffic patterns and the (slight) difficulty of negotiating the intersection.
Yet there is at least one prediction made by the theory of perfect competition that might be testable by a firm survey, as may be seen from Figure 3. Consider a firm which is operating in an environment of perfect competition, that is, it faces a flat demand curve. Suppose we approach the owner of this firm and ask him, "Would you like to sell more of your product for its current market price, if only you could?" The boss of this firm should definitely answer No. He is producing at a point where marginal costs are rising, and for any quantity above his current Q*, marginal costs are above the market price.
Probably there is a literature on how firms actually answer this question, but in any case many economists proceed on the (plausible) assumption that most firms would answer Yes. There is nothing paradoxical about that; it is merely inconsistent with perfect competition. It is consistent, instead, with a situation of what is often called by the oxymoron monopolistic competition, as shown in Figure 5:
In Figure 5, the individual firm faces a downward-sloping demand curve, like a monopolist. The firm chooses a price, P*, well above marginal cost. Here we must introduce the concept of marginal revenue, which was mentioned earlier but not explained. Marginal revenue represents the change in total revenues that occurs if a firm increases the quantity that it produces and sells, taking into account the need to reduce the price in order to sell a larger quantity. If the demand curve is flat there is no such need and marginal revenue is the same as price. But if the demand curve is downward-sloping, the marginal revenue curve slopes downward more steeply and lies below the demand curve. The firm maximizes profits by selecting Q* such that MR=MC.
Note that if the firm in Figure 5 were asked: "Would you like to sell more of your product for its current market price, if only you could?" he would answer Yes. Since the market price is well above his marginal cost, more sales at the same price would raise his profits. If most real bosses would answer Yes to the question, that suggests that the real world is better described by the monopolistic competition than by the perfect competition model.
It should be noted that the market in Figure 5 is characterized by large deadweight losses. It could produce much more than it is producing, and sell it for more than marginal cost, though it would reduce its own profits thereby. Perhaps oddly, the ideal solution might be to enable the monopolist, somehow, to engage in perfect price discrimination, charging different prices to different customers and particularly charging each customer his own willingness-to-pay. In that case, the monopolistically competitive firm would keep producing up to the point where P=MC, with large social gains, though none of these would be captured by consumers. Alternatively, the state might be tempted to intervene against "monopolistic practices" and force the firm to produce all units for which someone's willingness-to-pay was above the firm's marginal cost. But-- and this is one of Schumpeter's major points-- "monopolistic practices" can be socially beneficial. The most well-understood such case has to do with patent protection, as shown in Figure 6:
Figure 6 describes a firm which has an opportunity to innovate a new product. The cost If, having developed the product, the firm can act as a monopolist, it can recover profits worth area A. In that case, if A>B, as it is drawn to be in Figure 6, then the firm will find it worthwhile to develop the technology.
However, if there is free entry, the firm may not be able to charge P-M, the monopoly price, but may be forced by competition to reduce price to marginal cost. In that case, the technology will not be introduced.
The standard solution to this problem, of course, is to grant the innovating firm a patent or trademark or copyright or some other form of intellectual property in its novel product, granting it a temporary monopoly. Thus, in this special case, the case for allowing monopolistic practices is well-understood and widely recognized. What Schumpeter suggests, however, is that patents are only the tip of the iceberg, and that in conditions of the "gale of Creative Destruction," firms frequently need to engage in what appear to be monopolistic practices in order to protect themselves. Schumpeter argues that these practices can make the economy grow faster, and that "there is no more of a paradox in this than there is in saying that motorcars are traveling faster than they otherwise would because they are provided with brakes." (Schumpeter, 88)
Another example of monopolistic practices is that firms may deliberately incur large sunk costs in order to present a credible threat to fight entrants. Figure 7 shows how this strategy works:
In Figure 7, two versions of an "Entry Game" is represented in normal form. In each, there are two players, the Incumbent and the Entrant. Prior to game play the Incumbent is assumed to be a monopolist in the industry. The Entrant moves first and decides whether to enter (left) or stay out (right). The Incumbent then decides whether to collude by reducing production or to engage in rivalry by maintaining production.
The Incumbent, of course, wishes that the Entrant keep out, and it can, moreover, cause the Entrant to suffer losses should he enter. However, in the first (left-hand) version of the game, the Incumbent cannot credibly threaten to engage in rivalry, since he would in that case incur losses himself. The Entrant can foresee that if he plays Enter, the Incumbent will play Collude.
The second (right-hand) version of the game modifies the pay-offs on the assumption that the Incumbent chooses to engage in some kind of large capital investment. He incurs an unrecoverable fixed cost of 20, but thereafter is able to produce goods for a marginal cost three times lower than previously. From society's point of view, this investment does not appear to be very helpful overall. There is simply not enough demand for this large capital investment to do better than break even. However, having incurred these sunk costs, the monopolist now must produce at a relatively large scale in order to cover his fixed costs. If he reduces production and colludes, he incurs losses as large as if he engages in rivalrous competition. Foreseeing this, the Entrant will stay out. And this sheds light on the reason why the Incumbent is likely to engage in investment, namely, it enables him to pose a credible threat of fighting the Entrant.
Schumpeter reconceives competition in a way that departs quite sharply from the classical and neoclassical economists. This post is an effort to show why such a new departure in our understanding of competition is needed. In fact, although Schumpeter has been widely revered for his ideas about "creative destruction," the difficulty of formalizing this notion of competition (which Schumpeter himself does not even attempt to do) has limited its influence on academic economics. Yet there have been recurrent attempts to bring Schumpeter's ideas in. The "new growth theory" of Paul Romer and others pursues a research agenda closely related to that of Schumpeter. The idea of "contestable monopoly" that has become important in industrial organization is also anticipated in Schumpeter. Schumpeter points the way towards resolving certain deep tensions in economic thought going all the way back to Adam Smith and The Wealth of Nations, which have often caused economists to lose sight of Smith's fundamental insight about the connection between economic improvement and the division of labor.
Adam Smith's belief that division of labor raises productivity tempts us to formalize it using a downward-sloping supply curve, as in Figure 8:
Intuitively, Figure 8 is not hard to understand. As the size of the market grows, producers can benefit from growing economies of scale, and they sell their goods for a lower price. An increase in demand results in a larger quantity of a good sold, as well as a lower price: Q2>Q1 and P2<P1. This seems to be what's happening in the long run in the world economy-- we produce more, and lots of stuff becomes cheaper, at least in terms of labor if not of money-- and it's part of the story in particular success stories, such as that of East Asian countries, which became much more productive with the help of access to large world markets (especially the US).
But we must bear in mind that Figure 8 is fundamentally inconsistent with the assumptions of profit-maximizing firms and perfect competition. In order to justify a story like that conveyed in Figure 8, we must conceive of competition in a different way, such as that proposed by Schumpeter.
Recent Comments